1. Field of the Invention
The present invention relates to a semiconductor memory of a multi-bit type which stores and outputs a plurality of bits at the same time, which form a multi-bit data.
2. Description of the Related Art
A digital system having a semiconductor memory such as a DRAM (Dynamic Random-Access Memory) has the following three features to increase the speed of transferring data.
First, the semiconductor memory is a multi-bit type which can store and output the bits forming a multi (xc3x972n)-bit data (n is a natural number), at the same time.
Second, the semiconductor memory stores and outputs data in synchronism with an external clock signal supplied from a CPU (Central Processing Unit). The higher the frequency of the clock signal, the faster the memory (e.g., a SDRAM or a RDRAM) can store and output continuous data. Thus, the memory can serve to increase the speed of transferring data.
Third, the semiconductor memory incorporates a plurality of banks. The banks have identical elements each. The banks can store and output data, independently of one another. These measures taken, the time required to access the first data (known as xe2x80x9clatencyxe2x80x9d) is short, thus enhancing the speed of transferring data.
FIG. 1 shows the layout of a conventional semiconductor memory. This memory has all three features mentioned above.
The conventional semiconductor memory comprises a memory chip 10 and four banks 11-0 to 11-3 provided in the chip 10. Each of the banks 11-0 to 11-3 comprises a memory cell array and peripheral circuits such as a cell-array controller, a row address decoder, a column address decoder and a DQ buffer (i.e., a buffer provided in the input/output section of the bank).
The memory chip 10 has input/output region 12. Provided in the data I/O region 12 are a plurality of input/output (I/O) circuits. For example, 16 I/O circuits are provided if the memory chip 10 is designed to store and output 16 bits (i.e., 16-bit data, or 2-byte data) at the same time.
The memory chip 10 also has a data bus 13. The bus 13 extends between one block consisting of the first and second banks 11-0 and 11-1 and the other block consisting of the third and fourth banks 11-2 and 11-3. The data bus 13 is designed to transfer data (e.g., 16-bit data) between each bank and the data I/O region 12.
How the memory chip 10 stores and outputs multi-bit data will be explained.
First, one of the four banks 11-0 to 11-3 is selected. In the bank selected, the memory cell array is accessed on the basis of an address signal. As a result, 2n-bit (e.g., 16-bit data, or 2-byte data) is output from the bank. The 2n-bit data is supplied to the data I/O region 12. The I/O region 12 outputs the data. Thus, the 2n-bit data is output from the memory chip 10.
It is desired that the ratio of the region occupied by the data bus 13 to the all chip area be reduced as much as possible. In other words, the bus 13 needs to be made as thin as possible to decrease the chip area.
However, the greater the number of bits the chip 10 simultaneously store and outputs, the thicker the data bus 13, and hence the larger the region the bus 13 occupies. More specifically, as the number of bits which the memory chip 10 can simultaneously store and output increases (from 16 bits to 32 bits and hence to 64 bits), the area of the memory chip 10 inevitably increases.
The present invention has been made to solve the above-mentioned problem inherent in the conventional semiconductor memory. Its object is to provide a semiconductor memory of multi-bit, clock-synchronized and multi-bank type which can transfer data at high speed, without having its chip area increased.
To achieve the object, a semiconductor memory according to a first aspect of the invention comprises a memory chip and a plurality of banks arranged on the memory chip, for storing and outputting multi-bit data, independently of one another.
Each of the banks has a plurality of memory-cell blocks, at least one column decoder, a plurality of row decoders, a plurality of DQ buffers, and a a cell-array controller.
Each of the memory-cell blocks comprises two sub-blocks, sense amplifiers, word lines, data lines and column-selecting line. Each sub-block is composed of one memory cell array. The sense amplifiers are located between the two sub-blocks. The word lines, data lines and column-selecting lines are arranged on the memory cell arrays constituting the two sub-blocks. The memory-cell blocks are paced apart along columns of memory cells and the column-selecting lines and data lines. The sub-blocks are spaced apart also along the columns of memory cells.
The column decoder is located at at least a first end of every column of memory cells. It is connected to the column-selecting lines.
The row decoders are located at a first end of every row of memory cells along which the word line extend, and are connected to the word lines. Each row decoder is provided for one memory-cell block.
The DQ buffers are located at a second end of every row of memory cells. Each DQ buffer is provided for one memory-cell block;
The cell-array controller is located at a second end of very row of memory cells, for controlling the reading and writing of the multi-bit data.
The semiconductor memory further comprises a data input/output region and a data bus. The data region is provided on the memory chip, for receiving multi-bit data from an external device and outputting multi-bit data to an external device. The data bus is provided for the plurality of banks, extends parallel to the columns of memory cells, for transferring multi-bit data between the plurality of banks, one the one hand, and the data input/output region, on the other.
Each bank has local DQ-line pairs and global DQ-line pairs. Each DQ-line pair is provided between the two sub-blocks of one memory-cell block and extends parallel to the rows of memory cells. The global DQ-line pairs extend over the memory-cell blocks, along the columns of memory cells. They connect the local DQ-line pairs to the DQ buffers.
A semiconductor memory according to a second aspect of the invention comprises a memory chip and a plurality of main banks. The main banks are arranged on the memory chip, for storing and outputting multi-bit data, independently of one another. Each main bank is composed of a plurality of sub-banks.
Each of the sub-banks comprises a plurality of memory-cell blocks, at least one column decoder, a plurality of row decoders, a plurality of DQ buffers, and a cell-array controller.
Each of the memory-cell blocks comprises two sub-blocks, sense amplifiers, word lines, data lines and column-selecting line. Each of the sub-blocks is composed of one memory cell array. The sense amplifiers is located between the two sub-blocks. The word lines, data lines and column-selecting lines are arranged on the memory cell arrays constituting the two sub-blocks. The memory-cell blocks are spaced apart along columns of memory cells and the column-selecting lines and data lines. The sub-blocks are spaced apart also along the columns of memory cells.
The column decoder is located at at least a first end of every column of memory cells. It is connected to the column-selecting lines.
The row decoders are located at a first end of every row of memory cells along which the word line extend. They are connected to the word lines. Each row decoder is provided for one memory-cell block.
The DQ buffers are located at a second end of every row of memory cells. Each DQ buffer is provided for one memory-cell block.
The cell-array controller is located at a second end of very row of memory cells, for controlling the reading and writing of the multi-bit data.
The semiconductor memory further comprises a data/input region and a plurality of data buses. The data input/output region is provided on the memory chip, for receiving multi-bit data from an external device and outputting multi-bit data to an external device. The data buses are provided for at least two of the sub-banks and extend parallel to the columns of memory cells, for transferring multi-bit data between the sub-banks, one the one hand, and the data input/output region, on the other.
Each sub-bank has local DQ-line pairs and global DQ-line pairs. Each local DQ-line pair is provided between the two sub-blocks of one memory-cell block and extends parallel to the rows of memory cells. The global DQ-line pairs extend over the memory-cell blocks, along the columns of memory cells, and connect the local DQ-line pairs to the DQ buffers.
A semiconductor memory according to a third aspect of the invention comprises a test circuit, a memory cell array composed of a plurality of memory-cell blocks, data-writing means for writing bits of data simultaneously into memory cells of at least one of the memory-cell blocks, and a register for holding the data to be written into the at least one of the memory-cell blocks.
The test circuit comprises write/read means, comparing means and an output circuit. The write/read means is designed to write the data held in the register into the memory cells of the at least one of the memory-cell blocks and to read data from the memory cells. The comparing means compares the data held in the register with the data read from the memory cells to determine whether the semiconductor memory is flawless and for generating one-bit data representing whether the semiconductor memory is flawless. The output circuit is designed to output the one-bit data generated by the comparing means, from the semiconductor memory.
The test circuit further comprises latch means for holding the n-bit data generated by the comparing means, and switching means for supplying n bits of the data generated by the comparing means, sequentially to the output circuit when the comparing means determines that the semiconductor memory is defective.
The semiconductor memory having the test circuit is an n-bit type which can store and output n bits of data at the same time, which has n output pads for use in normal operating mode. In the test mode, one of the n output pads is connected to the test output circuit of the test circuit.
Additional objects and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. The objects and advantages of the invention may be realized and obtained by means of the instrumentalities and combinations particularly pointed out in the appended claims.